PresentationPDF Available

Regulations and ethics of AI applications in mental health - A complicated formula

Authors:

Abstract

Whilst the application of artificial intelligence promises immense potential benefits to virtually all disciplines and activities, it also raises new risks, in scope like in scale. This duality is especially relevant with health applications and more particularly mental health. Making decisions in regards to the application of AI in healthcare based on the evaluation of risks and benefits should be made according to ethics principles. - Practically, a complex framework of national and international regulatory bodies exists to establish and enforce regulations that define the conditions under which new technologies can be developed, validated, marketed and deployed for research, wellness or clinical use. - Our work identifies and defines the different modes under which current and emerging data-driven AI applications for mental health are regulated, and will assess how these modes address the main principles identified in the many AI ethics principle declarations published in the last 6 years by governments and international institutions. - More particularly, we evaluate how the concepts of trust, safety and transparency are embedded into these regulatory modes, whether these modes contribute to a more trusted adoption of AI tools into the clinician-driven decision making process, and the mechanism by which this adoption can impact the bias, accuracy, and precision of the decisions thus made. https://www.youtube.com/watch?v=_Z2FSAfm8e0&list=PL1wkAjrJw4HZ8ovyTJl0beDLNuCVrksxD
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
Regulations and ethics of AI
applications in mental health
A complicated formula
Pierre Lonchampt
Bahareh Nakisa, Steven Umbrello
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
Background
Ms. Physics and Eng., PhD Industrial engineering / Design research
Incl. Cognitive psychology
Engineering management / Medical devices
Regulatory Affairs and Quality Assurance
FDA, TGA, EU
HW / Systems
SW / AI
AI - Singularity - Tech Ethics
Intelligence - Conscience/Consciousness
Audio/Music
Mental Health
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
Motivation
● ATHT.ORG
○ NFP
2 Sides
Rejuvenate my research and education profile
Value Engineering - Co-evolution / Design
Opportunity for silo breaking
DIY (Mental) Health
Resolving cognitive dissonance
■ Values
Contributing back
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
Contents
Regulations for tech (incl. AI) in mental health
Who and What?
Link with Ethics
So what?
○ SWOT
Transparency and future research
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
Regulations? What (is the “thing” regulated …. or not!)
and by whom ?
The (medical) act
Intervention/therapy, diagnostic/test, education/prevention
■ Enhancement/modification
● “Wellness”
Self-practice and illegal practice
Research
The tech itself
HW, SW, Data
Codes, Models, Datasets
Case of “instructions” for building the tech
For apps:
The HW running the app
The distribution channel (Apple/Android store…)
The combination of tech and use:
“Intended use” (manufacturer’s label) vs. actual use
User = patient or professional?
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
Regulating “the act”
Professional boards
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
Regulating research
Ethics committees / IRBs…
Nuremberg to Helsinki
Differences “investigator initiated”
versus “sponsor initiated” trials.
Scientific publications vs closed-source
“commercial in confidence” regulatory
reports.
Data. HIPPA. GDPR
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
Regulating the tech (and its use)
Regulatory Bodies
Intended use and indications
for use
Pre market
Post Market
Medical device’ means any instrument, apparatus,
implement, machine, appliance, implant, reagent for in
vitro use, software, material or other similar or related
article, intended by the manufacturer to be used, alone or
in combination, for human beings, for one or more of the
specific medical purpose(s) of:
diagnosis, prevention, monitoring, treatment or alleviation of
disease,
diagnosis, monitoring, treatment, alleviation of or
compensation for an injury,
investigation, replacement, modification, or support of the
anatomy or of a physiological process,
supporting or sustaining life,
control of conception,
disinfection of medical devices
providing information by means of in vitro examination of
specimens derived from the human body;
Does not achieve its primary intended action by
pharmacological, immunological or metabolic means, in or
on the human body, but which may be assisted in its
intended function by such means. (Those are “drugs”)
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
Regulating the tech (and its use)
● Classification
Aka thresholds
Quality Management System
Safety and Effectiveness / Conformity
Principles
○ Equivalence
Risk Management
Risk - Benefit
Design controls
Inputs … Validation
Tech Standards
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
Regulating the tech (and its use) Harmonization
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
Other voices 1/4
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
Other voices 2/4
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
Other voices 3/4
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
Other voices 4/4
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
A complicated
formula
www.diapason.consulting/20210908
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
● Strength
● Weakness
● Opportunity
● Threat
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
SWOT
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
Risk - Benefit
Trying to balance Risks versus Benefits
SWOT
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
Risk - Benefit bias dimensional misconception
Trying to balance Risks versus Benefits
Promises
Harms
SWOT
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
SWOT
Discrete class and thresholds:
Historical legacy from a bureaucratic system
Responsible for chaotic behaviour and energy wasting navigation
Not intuitive (condom = critical care ventilator)
Barrier to entry - missing out on benefits
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
TRUST agents ? SWOT
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
Regulators to (only) enforce transparency? SWOT
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
SWOT
Silo Spectrum - continuum Silo
From experimental trials to standard of care
Ethics & values differences between
RCT (science)
Case Studies (conscience)
LMICs trials
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
SWOT
● Classification
Aka thresholds
Quality Management System
Safety and Effectiveness / Conformity
Principles as VALUES
○ Equivalence
Risk Management
Risk - Benefit
With transparency
Design controls
Inputs … Validation
Tech Standards
○ Yes...but
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
SWOT
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
SWOT
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
SWOT
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
SWOT
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
SWOT
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
SWOT
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
SWOT
Inspirations (to ‘cope’ with AI)
Science Arts
Explicit Implicit
Explainability Intuition
Societal introspection
Political correctness
Collective consciousness
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
SWOT
Innovation 1.0 Invention
Innovation 2.0 Collective / within the boundaries of incorporated entities (or gvt)
Innovation 3.0 Collective / without boundaries
Innovation 4.0 + AI
Hard science world
Concepts world
Mental health
Physical health
Neurology
Genomics
Atoms
Bits
Memetics?
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
Conclusion
Regulations need to evolve. To be ethical, they must:
Be able to compare fairly:
The risk of skin irritation from sensors
The risk of a hacker obtaining POTUS’s therapy notes
The risk of being ignored
Be based on an all-encompassing framework
Provide trust in technology
Empower patients and professionals to make informed decisions
Hypothesis/research questions
Transparency as a core value
OpenSource as a tool
Proactive rather than reactive compliance
We have the opportunity to DESIGN this regulatory approach
At the same time we are designing AI
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
https://www.linkedin.com/groups/13990791/
(Group Music, Health and Technology)
www.atht.org
@TranspaTech
www.diapason.consulting
@DiapasonConsult
www.lilialaska.com
Melbourne 13 Feb 2022
Sydney 19 Feb 2022
w
ww.sofrenchysochic.com
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
For questions…….
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
SWOT
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
There cannot be trust without transparency
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
Evolution of the roles of the regulators?
Is it safe?
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
Evolution of the roles of the regulators?
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
2021 Symposium on Artificial Intelligence in Mental Health - Pierre Lonchampt et. al. WWW.DIAPASON.CONSULTING www.ATHT.ORG
Attribution-ShareAlike 4.0 International
Required Evolution for the Regulatory framework:
(ref www.diapason.consulting/20210908
Lean and Agile, Ad-Hoc (Bottom-up)
Embedded and Participative, Incl. Ethicist
Culture of ETHICS
● Global
Continuous - Efficient
● Inclusive
● Transparent
Hypothesis:
‘Proactive transparent accountability’ is better than ‘benchmark
compliance’
ResearchGate has not been able to resolve any citations for this publication.
ResearchGate has not been able to resolve any references for this publication.